Since the international standardization in video coding methods had recently been making progress, many video contents are often used in our everyday life. Moreover, there are some problems with the methods of application in multi-quality videos due to the growth of mobile devices, e.g., smart-phones and tablets, and
the image resolution enhancement of TV images, e.g., HDTV, 4K-TV and 8K-TV. In particular, the application of scalable video coding (SVC) has attracted the attention. Meanwhile, SVC has the problem of the reduction of coding quality due to the increase in hierarchical header information. Then, we propose a reduction scheme of the coded bits using a prediction method of quantization parameters in SVC. In this study, we describe the computed parameters and consider the improvement of the coding efficiency.
Digital signage is installed at various places and plays the role of key advertisement medium regardless of the
location in everyday life. Since digital signage is a kind of push-type medium like television or radio, it is expected that the
digital signage can be useful source of disaster information at the place where one is staying if we can utilize it to transmit this
information. We propose a efficient method for transmitting disaster information through local-government-managed digital
signage of which it can easily control the sending contents. We also report the result of an experiment which was carried out
for evaluating the effectiveness of the proposed method.
The Great East Japan Earthquake that occurred on March 11, 2011 brought serious damage in the Tohoku district in Japan. There have been many examples of utilization of satellite image analysis to grasp the situation of damage. However, henceforth, it is more important to monitor how the destructed environment will recover for a long time. From such a background, our research group is working on a project such as “A project on the recovery monitoring of the destructed environment in the disaster area by satellite data analysis and field work and the practice of earth environmental education.” The main purposes of the project are to monitor the status of the environment recovery from both aspects of satellite image analysis and fieldwork with high school students in the disaster area, and to practice earth environmental education for young people through such work. Moreover, to construct Web-based system to present the status of the environment recovery and the reconstruction of the disaster area clearly and to transmit this information all over the world are also purposes of the project. In this paper, we report on the implementation of a Web-based map to present the status of the recovery from the Great East Japan Earthquake by using both the data obtained by the fieldwork conducted by the project members and the National Land Numerical Information provided by the MLIT (Ministry of Land, Infrastructure, Transport and Tourism).
This paper describes a characteristic M2M and IoT technology which has been realized through international joint effort of Hungarian company ITware with Japanese partner FB Triangle. ThebasicM2MsystemnamedKojimori has been applied to various application area, and application development editor named APPaware has also been applied to the management and operation for IoT environment of Kojimori, with interface devices of smartphones and tablet PCs. Then applications to IoT area have been studied for two categories. One is the moving object sensing with GIS application while another is ontological application for milling machine operation skill transfer through ontology based operation manuals.
In the plastic surgery operation to the patient who lost a part of face (an eye, an ear, an external nose, a cheek, etc.) by an external injury etc., regularity of the form of the part reconstructed by operation is very important for the patient’s QOL. However, the curve of outline of the form is decided by the plastic surgeon's sensitivity, and the problem that the regularity changes greatly with the the plastic surgeons’ levels of skill and plastic senses has occurred now. Then, in collaboration with a plastic surgeon, I have analyzed the curves of outlines (sections) of the forms (an eye, an ear, an external nose, a philtrum), and classified those curves in the latest several years. Moreover, I introduce the examples of development the templates using the result and application to the actual plastic surgery operation.
Facial expressions play an important role in all facets of human communication, even in human-machine communications. Since the facial expressions displayed on an avatar face are interpreted more accurately in 3D than in 2D, we must properly transform the 3D shape of its original straight face to create a 3D avatar face that displays various expressions. Morphable 3D face model was introduced by learning from many 3D faces that display various expressions, with which each transformation of the 3D shape is described by changing just a few parameters. Moreover, an impression on personal attributes can be changed by controlling these parameters. In our previous works such parameter transformation suitable to generate facial expressions was proposed by introducing impression transfer vector, which they uniformly obtained for any input face using linear discriminant analysis. We proposed a new method for individually obtaining the transformation of the parameters for each input face on the basis of the positional relationships between the parameters of the input face and the discrimination boundaries defined in the parametric space by Support Vector Machine learning.
This work analyzed gaze distribution data on facial images obtained from observing subjects by using an eye-tracking system while they made different impression judgments: this result might depend on the dimension of the personality impression being observed. In particular, we conducted experiments on gaze distribution analysis in terms of the following two issues: whether the gaze distribution pattern differs by the contents of the impression judgment and whether it is influenced by facial makeup.
Users of smartphones and/or tablet computers browse and download confidential document files routinely. Therefore, the higher security level is needed for smartphones and tablet computers than conventional mobile phones (feature phones). From this kind of background, Takahashi and Uchida proposed an image-based user authentication method for touch screen devices by using the latest image shot by user oneself as the pass-image. The authentication method proposed by Takahashi and Uchida has resistance to smudge attacks, one of the most serious threats for touch screen devices. However, the volume of the password space of their method is smaller than the one of 4 digit PIN codes. Therefore, in this paper, we propose SWIPASS, an image-based user authentication method for touch screen devices that has both resistance to smudge attacks and high security strength, by improving the method proposed by Takahashi and Uchida. The volume of the password space of SWIPASS is 41,472. Then the security strength of SWIPASS is superior to the one of 4 digit PIN codes. Moreover, we conduct several experiments assuming shoulder surfing attacks to evaluate the resistance against shoulder surfing attacks.
In recent year, device operation by eyes input appear. The device has merit which can be operation without hand. However, user can't be pointing stable by eyes input. Therefore, it is difficult to perform stable target selection. In this paper，gaze point is regarded as which has spatial spread. Then, considered the past of gaze point. We propose a method of select Icon which is the most of gaze likelihood. Thereby, target selection is improved extremely 32.7% to 76.2%.
It is impotrant in industrial design to generate high-level aesthetic curves. In general, high-level aesthetic curves is defined that curvature change is monotonous or the slope of logarithmic curvature graph is approximated straight line. However, the cubic polynomial parametric curves often used in a CAD system has the complicated relation between a control point and curvature change. In this study, We focus on the association of the parametric cubic curves and the aesthetic curves. The aesthetic curves has the high-level character and the monotonic curvature. We aim at the analysis in consideration of the aesthetic factor. Our study will lead to develop a new control system.
Our multi-directional switching median filter is one of several types of filters which detect noisy pixels in an image and denoise them. A threshold setting is necessary with the optimal value depending on the quantity of each image's edge. We focus on this point and propose an improved method that divides an image into segments and, for each one, sets a threshold proportional to the quantity of the edge. Generally, SIDBA images are used as test images. Compared with current large-scale images, test images are smaller in the number of pixels and their subjects are outdated. Therefore, we evaluated the denoising performance of our filter using more realistic large-scale image and this paper reports our results.
On the field of stem-cell research and drug discovery, to extract cell region from microscopy images and to perform accurate identification and screening of cells are needed. Since most of these works are done by hand, it takes a lot of time and cost. Moreover, there are done by multiple people, objectivity would be lacked. Therefore, automated processing by computer have been studied. This paper examines texture analysis and segmentation by Chan-Vese level set method of cell images where cells are crowded, and evaluate their results cell division identification point of view.
In this paper, we propose a method for 4D cell image segmentation and visualization. We divide the domain by texture analysis and a statistical model．And we get the shape of a fluent outline by the matching of the shape next. We perform the three-dimensional expression of the cell image by the volume rendering. Experimental result shows that the proposed method can segmentation the cell, and visualize 3d structure inner cell.
During the cell division a number of biochemical reactions are taken place inside the cell. As a result, we observe the motion in the materials inside the cellular membrane. Most of the cellular components are residing in the cytoplasm. Cytoplasmic streaming helps to transport the materials between organelles, as well as between cells. In fact, cytoplasmic streaming is a motion in 3D space. This motion can be observed by using 3D time-lapse image dataset. However, it is very difficult to determine the 3D motions subjectively. Subjective analysis is very important for describe the biological properties in the cytoplasmic streaming as well as verification of objective assessments. This paper illustrates implementation details of visualization tool that shows 3D motion vectors of cytoplasmic motion. In addition, we describe its usability for advanced biological investigations.
Magnetic Resonance Imaging (MRI) movie of the tongue motion was acquired for Japanese “a”, “ka”, and “ga” columns. We explain this method and a few problems the speech researchers use image processing tools. We tested more phoneme patterns, /aeiueoao/, /kakekikukekokako/ and /gagegigugegogago/. The “synchronized sampling method” was used for the acquisition of the speech, and the subjects repeated the same task by listening to the guide tones. (1) The images of the same phonemes were compared, and (2) The kymograph was made. The horizontal-direction corresponds to the time, and the vertical direction implies the distance to the soft palate and the speech organ. The kymograph showed different tendencies from the previous speech research. MRI artifact should be also taken into account carefully.
As an extension of log-aesthetic curves, this paper presents a method for drawing curves with quadratic logarithmic curvature graphs and clarifies the fundamental characteristics, such as the arc length of point at zero or infinite curvatures. Let α and γ be the linear and quadratic coefficients of logarithmic curvature graphs. Curves with γ = 0 are planar log-aesthetic curves. For some αs and γ, we show how the curvature plots and curve shapes of the proposed curve change as well as the fundamental characteristics. In comparison with log-aesthetic curves, the proposed curves have following advantages: additional parameter γ for controlling the curve and curves with γ‹ 0 have inflection points without depending on the value of α.
It is hard to detect a partially occluded and weak perspective projected maker in vision-based AR(Augmented Reality). Moreover, a CG(Computer Graphics) image should be superimposed into an AR marker even if the marker is composed by uneven polygonal surfaces. First, the marker is detect a partially occluded by SIFT(Scale-Invariant Feature Second, weak affine projected regions are extracted with spatial density of SIFT key-points. Finally, the three-dimensional Transform). posture of plane of the weak affine region are Extracted by Structure from Motion.
This paper proposes an efficient importance sampling method of virtual point lights (VPLs) for many-light rendering by constructing probability distributions proportional to the contributions of VPLs. In many-light rendering methods, direct illumination and indirect illumination incident onto shading points are approximated by the contributions of VPLs. The proposed method employs an importance caching method to construct probability distributions of sampling VPLs. The importance caching method records the contributions of VPLs at each cache point to construct probability distributions. The importance caching method, however, generates cache points randomly, which makes it difficult to obtain probability distributions proportional to the contributions of VPLs at shading points where the variations of contributions from VPLs to nearby cache points are large. To address this problem, the proposed method distributes cache points adaptively for such regions to construct probability distributions proportional to contributions of VPLs, resulting in rendering less noise images with the same number of VPLs.
For creating new tactics of sports like volleyball, the analysis of player motion in real games becomes more and more important. However, since motion data needed for the analysis is captured by human observation currently, an automatic capturing system from video camera is highly expected to gather many useful data easily. This paper proposes an action detection algorithm of volleyball players using motion features based on clustering and aggregation of body trajectories. Since the body trajectories of arms and legs are similar, the clustering utilizes shape, location and density of their trajectories. Furthermore, the clustered feature values are aggregated by means of their mean and variance. Experimental results by using the motion detection system based on the proposed algorithm show that it averagely attains 0.9539 AUC of the ROC curve for the detection of four basic motions (block, receive, spike and toss) from the volleyball game video captured by high-definition cameras. This is 0.014775 higher than conventional methods.
We propose a dense motion analysis method for 3D ultrasound images. We locate a lot of lattice points on the 3D ultrasound image, and analyze their motion. Our method consists of two processings. One is the tracking of the lattice points based on an optical flow algorithm in a framework of spring-model. The other the detection of the lattice points based on textural information. The detection results correct the tracking errors, and hence, the accumulative error is suppressed. Furthermore, we only use the detection results that have high confidence value, and hence, the false corrections of the tracking errors drastically decrease. We evaluated our method using a sequence of artificial 3D ultrasound images up to 5 minutes. The average and maximum error of our proposed method achieved the best performance in the conventional methods.
The purpose of in this study is to detect fine movements of human, by using edge direction variations. Edge direction does not easily change by noise such as illumination variation. Also, the edge direction can vary greatly even if the person is slightly moved body. Therefore, it is presumed that the edge direction difference is suitable for the detection of fine movements of the person. First, the detection of the moving object is implemented using an edge direction difference between
adjacent frames, and analyzed its properties and trends.